[Geezer (OP)]

I have an old metal turning lathe. It's about ten years older than me! I don't know if you are familiar with how screw threads are cut on a lathe, but it's really fairly simple.

The cutting tool holder can be engaged on to a "leadscrew" that runs along the length of the lathe. There is a system of gears that the operator can set up so that, as the item being turned rotates with the headstock, the cutting tool will traverse along the bed of the lathe and cut a helix into the piece being turned.

The gear ratio that the operator selects between the headstock and the leadscrew determines the pitch (distance between threads) of the helical screw thread being cut.

The pitch of the leadscrew on my old lathe happens to be one eighth of an inch - or, there are eight threads per inch if you prefer. If I set up the gearing with a 1 to 1 ratio between the headstock and the leadscrew, the thread I cut will be the same as the leadscrew, or 8 threads per inch.

I have a set of gears that let me select a wide range of ratios so that I can cut almost any practical number of threads per inch (TPI).

Here is a list of the available gears and the number of teeth on each gear.
16 24 28 32 32 36 40 44 46 48 52 54 56 60 80.

There are also a couple of compound gears (that's two gears mounted on a common axis) which have ratios of 3:1 and 4:1, but I would not bother with them at the moment too much.

To cut metric threads with this lathe (that's thread pitches that are mostly multiples of one tenth of a millimeter (e.g, 0.5, 0.6, 0.8, 1.0, 1.5 etc) it's possible to obtain a transposing compound gear with 100 and 127 teeth. This provides a precise conversion from inch to metric. However, they are rather expensive because they are large gears.

I am able to cut metric threads with the addition of a single fairly small gear that I made, with almost the same precision. It's certainly precise enough for anything I'm doing.

How many teeth do you think are on that gear, and what precision do I get? (I suppose that should really be how imprecise is it?)

[SeanB]

20 and 25 teeth, would give a reasonable cut, although you would need to relax the tolerances slightly, and would not be able to use a long nut as it would bind somewhat. Good enough for most applications, and certainly as good as any modern lathe. Just remember that you would need to use a cutting tool that has the metric angles, as there are so many imperial angles that define screw threads.

[Bored chemist]

I realise it's not the answer yo are looking for, but can you get a metric leadscrew?

[Geezer (OP)]

I realise it's not the answer yo are looking for, but can you get a metric leadscrew?

That would be even more expensive than the transposing gears I mentioned. Also, it's not really very practical to do that. I need to be able to cut imperial and metric threads, and it's quite a lengthy operation to replace the leadscrew.

[Geezer (OP)]

20 and 25 teeth, would give a reasonable cut, although you would need to relax the tolerances slightly, and would not be able to use a long nut as it would bind somewhat. Good enough for most applications, and certainly as good as any modern lathe. Just remember that you would need to use a cutting tool that has the metric angles, as there are so many imperial angles that define screw threads.

Thanks Sean. Actually, I am able to do a lot better than the results I would get with 20 and 25.

If you had not already guessed, this is really a math puzzle, although the story about my old lathe is quite true. I'm interested to find out if anyone can produce a solution for the optimum single gear solution mathematically. I don't have one. I pretty much stumbled on the answer by accident! There may even be a better answer (closer approximation) than my answer.

[maffsolo]

http://www.lathes.co.uk/atlas/
Investigate this company's machine
It is capable to cut US standards, Metric, Whitworth.
Its gear guide is in the inner face of the head stock safety cover.

As cutting threads usually the index is your timer match for the carrage and lead screw.

I think my dad needed to modify this process for the Whitworth screw cut.
Keeping the carrage always engaged with the lead screw and put an electrical switch to reverse the rotation of the motor for the carrage return.

I think that the lead screw had dome play in it that could not be adjusted not sure.

But if you can get a hold of this gearing index guide, it should be able to help you adapt and translate to any ratio combination.

Machine Shop Math

I will try to find the guide online

 [ Invalid Attachment ]

[Geezer (OP)]

Thanks Maffsolo.

Yes, I'm familiar with that website. It has a lot of good information. You'll find my lathe there too. It's a South Bend 9 inch, model C.

BTW, I know about the technique of keeping the half-nuts engaged and reversing the lathe to cut certain threads, but I should warn everyone that it can be very dangerous. The chuck on older lathes is usually screwed on to the headstock. When you start the lathe in reverse, the chuck can actually unscrew itself from the headstock.

[maffsolo]

Model A horizontal drive

Draw bar locking headstock with collets is nice

 [ Invalid Attachment ]

This isn't your picture is it??? OOps no it  can't be you have a beard sorry.

[Geezer (OP)]

The S.B. Model C does not have a gearbox for the leadscrew (the A and B do). It has no power feed either, wich can be a bit of a pain!

I think it's the sudden start in reverse that makes the chuck unwind, although some machines have a grub screw to prevent that from happening. Also, it may not be a great idea to run lathes with plain metal bearings in reverse. I've noticed it sometimes seems to disturb the "set" of the soft metal in the bearing.

BTW, there is another trick for timing the engagements of the half-nuts on the leadscrew. I have a buzzer on my lathe that tells me when to throw the lever that engages them on to the leadscrew..

Never saw a "hendy" though.

[maffsolo]

Model C ok

 [ Invalid Attachment ]
 [ Invalid Attachment ]

Support the efort?

[maffsolo]

The S.B. Model C does not have a gearbox for the leadscrew (the A and B do). It has no power feed either, wich can be a bit of a pain!

I think it's the sudden start in reverse that makes the chuck unwind, although some machines have a grub screw to prevent that from happening. Also, it may not be a great idea to run lathes with plain metal bearings in reverse. I've noticed it sometimes seems to disturb the "set" of the soft metal in the bearing.

BTW, there is another trick for timing the engagements of the half-nuts on the leadscrew. I have a buzzer on my lathe that tells me when to throw the lever that engages them on to the leadscrew..

Never saw a "hendy" though.

Sorry for that geezer I read that post of mine and scubed it I had to much non scence there.

Buzzer thats cool I always was in suspence watching the small passive index wheel driven by the lead screw.
I always had to mis it and wait for the next available window of opportunity. I forgot if it was every half rev or the full rev

Make logical scence about that reverse. We never needed to do it but once. I needed threads made to accomadate a timing degree wheel on a  70 or 71 Triumph motorcycle it mounted on the stator magnet. Used to do dry engine timing.

[maffsolo]

Geezer here is a thread chart for an Atlas lathe
maybe looking and comparing the South bend layoutMaybe it will help with a gear ratio?

http://www.machinistweb.com/10-inch%20F-Series%20Parts%20List(2).pdf

[Geezer (OP)]

Thanks Maffsolo. I was trying to work out if the metric precision on the Atlas was better or worse than my method, but I keep messing up the calculation!

Anyway, according to my calculations, using my single additional gear I can cut a one millimeter pitch thread with an error of less than 0.0005 mm per pitch. That's less than 0.5 mm per meter, which is probably good enough for most of the things I do, except manufacturing leadscrews of course.

[maffsolo]

Thanks Maffsolo. I was trying to work out if the metric precision on the Atlas was better or worse than my method, but I keep messing up the calculation!

Anyway, according to my calculations, using my single additional gear I can cut a one millimeter pitch thread with an error of less than 0.0005 mm per pitch. That's less than 0.5 mm per meter, which is probably good enough for most of the things I do, except manufacturing leadscrews of course.

Glad to help Geezer.
My dad had a good handle on machine work.
I could never understand the gear layout on the Atlas Lathe. He would set that lathe up in less then 15 min. But that is then.

[Geezer (OP)]

I managed to figure out the Atlas threading chart.

It's not quite perfect either. A nominal 1 mm pitch will actually have a pitch of 1.00061, which is a slightly greater error than my arrangement on the SB, but it's still quite close. Also, it looks as though there are quite a few metric threads it won't cut, although it does seem to cover the most common ones.

BTW, the numbers I'm quoting assume the leadscrew has exactly 8 threads per inch, with zero error, which is highly unlikely! I don't know what sort of tolerance they allowed over the length of the leadscrew on these machines. The error from the leadscrew itself could conceivably be greater than the error introduced by the slightly imprecise gearing, or it it might even tend to cancel some of it out.

[maffsolo]

I was trying to find out what tread was on the lead screw for the Atlas,

This may relieve a question in your mind....
I Googles this phrase:
'number of threads per inch on a lead screw for an atlas lathe'

"Search ResultsAtlas - lathes In the USA large numbers of both the 10-inch 6-inch Atlas lathe were sold by Sear, .... housings: the "Horizontal" bolted to the bench at the back of the lathe (and made the whole .... Leadscrew: 3/4" diameter., 8 Acme threads per inch ...
www.lathes.co.uk/atlas/ - Cached - Similar"

Unless you need to make threads with MS specifications
In calculating the margin of error, going by what you calculated,
 Your homemade getup is far more  superior.
 

[Geezer (OP)]

Maff,

Yes. The Atlas uses 8 threads per inch, or 1/8 th inch pitch if you prefer. I was able to confirm that by looking at the gear setup for inch threading.

The error difference between the Atlas metric setup and my method is really very small. Mine is only slightly more accurate. Either setup is quite good enough for nearly any application. I think the only time the error introduced by either of these approximations would create a problem would be if it was used to produce a micrometer thread, or a calibrated leadscrew.

[soirish]

Perhaps you cut yourself a 47-tooth gear?  The ratio 47/60 is 0.78333 while the standard 100/127 is 0.7874, both rounded to four decimal places.  On an 8 tpi leadscrew, one turn of the screw produces 0.125" of travel, so the above ratios would result in 0.0979" and 0.0984" of travel respectively, a difference of roughly 0.0005" ("half a thou").

[soirish]

Sorry for leaving this part out in the original reply:  I arrived at this answer by using the standard change gear ratio of 100/127 by various gears available to you from your list and rounding the answer to achieve a gear size that best fit your requirements.
(100/127) * 60 = 47.244... for a 47-tooth gear.  Dividing 47 by the trial gear of 60 results in a ratio very near 100/127, as I described in the first part of the reply.

[soirish]

Trying this again, I find that a 41-tooth gear is better:  41/52 results in a ratio of 0.7884, which itself results in a travel of 0.125" * 41/52 = 0.09855..." or 2.503mm.